Profiling the Immune Response to Periprosthetic Joint Infection and Non-Infectious Arthroplasty Failure

Arthroplasty failure is a major complication of joint replacement surgery. It can be caused by periprosthetic joint infection (PJI) or non-infectious etiologies, and often requires surgical intervention and (in select scenarios) resection and reimplantation of implanted devices. Fast and accurate diagnosis of PJI and non-infectious arthroplasty failure (NIAF) is critical to direct medical and surgical treatment; differentiation of PJI from NIAF may, however, be unclear in some cases. Traditional culture, nucleic acid amplification tests, metagenomic, and metatranscriptomic techniques for microbial detection have had success in differentiating the two entities, although microbiologically negative apparent PJI remains a challenge. Single host biomarkers or, alternatively, more advanced immune response profiling-based approaches may be applied to differentiate PJI from NIAF, overcoming limitations of microbial-based detection methods and possibly, especially with newer approaches, augmenting them. In this review, current approaches to arthroplasty failure diagnosis are briefly overviewed, followed by a review of host-based approaches for differentiation of PJI from NIAF, including exciting futuristic combinational multi-omics methodologies that may both detect pathogens and assess biological responses, illuminating causes of arthroplasty failure.


Total Joint Arthroplasty Failure
Total joint arthroplasty is a common restorative surgery. In the United States (US), more than 1.5 million people undergo total hip arthroplasty (THA) or total knee arthroplasty (TKA) yearly and it is anticipated that numbers of arthroplasties will rise due to the aging population, high rates of obesity, and physical activity throughout the lifespan, including in later years. It has been predicted that THA and TKA procedures will grow by 137% and 601%, respectively, in the US between 2005 and 2030, resulting in more than 5 million primary THAs and TKAs and an estimated US$1.85 billion in annual hospital-related costs by 2040 [1][2][3][4][5]. A study of over 1.5 million primary TKA or THA patients from the US National Inpatient Sample found the mean age of patients undergoing primary TKA to have decreased from 68 years in 2001 to 66 years in 2011, while the mean age of those undergoing primary THA decreased from 66 years in 2001 to 65 years in 2011. In total, 64% and 56% of those in the TKA and THA groups, respectively, were women [6].
There are several reasons patients undergo joint replacement surgery. Most commonly, the procedure is in response to symptomatic osteoarthritis, followed by inflammatory arthritides, such as rheumatoid arthritis, or joint damage due to injury, tumors, or osteoporosis; regardless of underlying disease, these conditions result in pain, loss of joint mobility, and/or an overall decrease in quality of life [7][8][9][10]. In most cases, joint replacement provides dramatic pain relief and restoration of joint function, although 1 to 3% of

Periprosthetic Joint Infection
PJI occurs in~1 to 3% of patients undergoing primary total joint arthroplasty and makes up 20% to 50% of arthroplasty failures [9,13,16,[18][19][20][21]. In 1995, the National Institutes of Health (NIH) Consensus Development Panel on Total Hip Replacement described PJI as a "devastating complication" that is "challenging" to diagnose [10]; antimicrobial agents plus surgery are needed to treat it. The type of surgical intervention is based on the infecting organism(s), timing and duration of infection, and clinical presentation. For acute PJI, joint irrigation, followed by debridement, long-term antimicrobials, and implant retention (DAIR), is typical, costing around US$67,000 [22]. For non-acute cases, resection of components is characteristically necessary, either as one-or two-stage procedures. Twostage exchange arthroplasties typically cost around US$114,000, though costs vary [9,[23][24][25][26]. In all, management of PJI costs hospitals~5-fold more than uncomplicated hip arthroplasty, totaling approximately US$2 billion per year, not including non-medical costs [5,27,28]. This cost is in addition to the individual patient burden, often including devastating morbidity, expense, impairment of quality of life, and potential loss of implanted devices, or even limbs, in extreme circumstances [9,29]. There are several risk factors associated with PJI, including obesity, smoking, and immune-disrupting disorders and their treatments, such as diabetes mellitus and rheumatoid arthritis. Men have been observed to be more prone to infection than women, although the biological underpinning for this observation is unknown [30][31][32][33][34][35][36][37][38][39][40][41].
PJI is typically caused by the formation of bacterial biofilms on device surfaces and in the surrounding tissues; biofilm formation involves the production of extracellular polymeric substances and immune-modulating products that protect microorganisms from antimicrobial agents and the host immune response [42][43][44][45][46]. Staphylococci, primarily Staphylococcus epidermidis and Staphylococcus aureus, are the most common causes, followed by polymicrobial infections, streptococci, anerobic bacteria, aerobic Gram-negative bacilli, and enterococci. In rare cases, other bacterial types, or even fungi, cause PJI ( Figure 1) [9,[47][48][49][50][51][52]. Culture-negative infections make up 6 to 15% of PJI cases, although rates up to 42% have been reported. Culture-negative PJI is particularly challenging due to the difficulty in selecting a treatment regimen for an unknown causative organism (which may be "over" or "under" treated) or even in classifying the joint as infected in the first place [47,[53][54][55].

Current Arthroplasty Failure Diagnostic Techniques
Despite the health and financial impact of arthroplasty failure, there are no universally accepted clinical definitions or diagnostic criteria for PJI and NIAF. Between 2001 Antibiotics 2023, 12, 296 4 of 26 and 2021, there was an~30-fold increase in PubMed yearly publications for all PJI and those specifically related to PJI (Figure 3). The influx of PJI-related content and rising PJI numbers have led to several organizations, such as the Infectious Diseases Society of America (IDSA) and the Musculoskeletal Infection Society (MSIS) in 2011, International Consensus on Orthopedic Infections Meetings in 2013 and 2018, and the European Bone and Joint Infection Society (EBJIS) in 2021, to propose diagnostic guidelines, although there is constant evolution and refinement as a result of new knowledge and improving diagnostic approaches; a global consensus definition of PJI has yet to be reached [25,[65][66][67][68][69]. It has been recently argued that while complex, multi-step approaches to PJI diagnosis may be useful in research settings, a single accurate differentiative assay would be most helpful in clinical practice [70]. In addition to distinguishing PJI from NIAF, an important consideration is a need to define the microbial etiology of PJI in the infected cases. Whether used in combination or as individual analyses, current diagnostic assays are primarily either microbial-based, such as traditional bacterial culture and molecular analyses, or host-based, such as histopathological evaluation and measurement of single host-based biomarkers. While able to discern PJI from NIAF in some cases, these techniques come with limitations.

Current Arthroplasty Failure Diagnostic Techniques
Despite the health and financial impact of arthroplasty failure, there are no universally accepted clinical definitions or diagnostic criteria for PJI and NIAF. Between 2001 and 2021, there was an ~30-fold increase in PubMed yearly publications for all PJI and those specifically related to PJI (Figure 3). The influx of PJI-related content and rising PJI numbers have led to several organizations, such as the Infectious Diseases Society of America (IDSA) and the Musculoskeletal Infection Society (MSIS) in 2011, International Consensus on Orthopedic Infections Meetings in 2013 and 2018, and the European Bone and Joint Infection Society (EBJIS) in 2021, to propose diagnostic guidelines, although there is constant evolution and refinement as a result of new knowledge and improving diagnostic approaches; a global consensus definition of PJI has yet to be reached [25,[65][66][67][68][69]. It has been recently argued that while complex, multi-step approaches to PJI diagnosis may be useful in research settings, a single accurate differentiative assay would be most helpful in clinical practice [70]. In addition to distinguishing PJI from NIAF, an important consideration is a need to define the microbial etiology of PJI in the infected cases. Whether used in combination or as individual analyses, current diagnostic assays are primarily either microbial-based, such as traditional bacterial culture and molecular analyses, or host-based, such as histopathological evaluation and measurement of single hostbased biomarkers. While able to discern PJI from NIAF in some cases, these techniques come with limitations.

Microbial-Based Diagnostic Techniques
Bacterial culture and molecular assays are traditionally used as microbial-based techniques for PJI diagnosis and pathogen identification [71]. Traditional bacterial culture consists of harvesting patient samples, such as synovial fluid, sonicate fluid (i.e., fluid generated from sonication of resected implants), and/or periprosthetic tissue, and inoculating them into or onto culture media, which are then incubated to assess for microbial growth. If microbial growth is detected, that growth is identified by additional analysis, such as matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry (MS) [71][72][73][74]. While bacterial culture remains the clinical gold standard for PJI diagnosis and allows recovery of microorganisms for antimicrobial susceptibility testing, it has limitations. Inherently, culture relies on growth in or on culture media such that inadequate growth conditions or low bacterial inocula can lead to negative results; further, culture is affected by antimicrobial usage before sampling (which is common). Due to these

Microbial-Based Diagnostic Techniques
Bacterial culture and molecular assays are traditionally used as microbial-based techniques for PJI diagnosis and pathogen identification [71]. Traditional bacterial culture consists of harvesting patient samples, such as synovial fluid, sonicate fluid (i.e., fluid generated from sonication of resected implants), and/or periprosthetic tissue, and inoculating them into or onto culture media, which are then incubated to assess for microbial growth. If microbial growth is detected, that growth is identified by additional analysis, such as matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry (MS) [71][72][73][74]. While bacterial culture remains the clinical gold standard for PJI diagnosis and allows recovery of microorganisms for antimicrobial susceptibility testing, it has limitations. Inherently, culture relies on growth in or on culture media such that inadequate growth conditions or low bacterial inocula can lead to negative results; further, culture is affected by antimicrobial usage before sampling (which is common). Due to these limitations, culture-based techniques may have imperfect sensitivity, even when infection is present. The type of sample analyzed also affects sensitivity. For example, periprosthetic tissue culture is generally less sensitive than sonicate fluid culture [73]. Contamination by extraneous microbiota may be a challenge with culture-based approaches, rendering determination as to whether isolated microorganisms are pathogens or contaminants difficult based on identity alone (e.g., S. epidermidis) in some cases. This can be overcome by culturing multiple samples (e.g., periprosthetic tissues) from each patient; recovery of the same species from more than one sample typically implies that it is a cause of infection. Time to detection is another limitation of culture-based methods; for example, anaerobic bacterial cultures are routinely incubated for 14 days prior to being reported as negative [9,75]. A 2022 study of 536 PJI patients found that the median time-to-positivity for all positive cultures was almost 3.5 days, although this was dependent on the microbial species, with S. aureus growing in the shortest mean time (1 day) and Cutibacterium acnes in the longest mean time (almost 7 days). Sample type also impacted time to results, with synovial fluid having the shortest mean time-to-positivity, followed by periprosthetic soft tissues [75].
Molecular assays, such as nucleic acid amplification tests (typically, polymerase chain reaction (PCR) assays) and those based on microbial sequencing, are increasingly used for PJI diagnosis and pathogen identification [9,71,[76][77][78][79]. As molecular techniques detect microbial components and not living bacteria, they may theoretically have higher sensitivities than culture-based assays. This increased sensitivity may come with drawbacks. Similar to culture-based assays, molecular techniques may be limited by low bacterial abundance and prior antimicrobial usage. Detection of contaminants or otherwise clinically insignificant bacterial components is a limitation of molecular-based techniques. Because of the sensitivity of these techniques, microbial elements from sample contamination during harvest and/or processing or left over from previous infections may be detected, leading to erroneous results. Molecular techniques may be more expensive and have longer turnaround times than culture-based techniques, although this is not always the case. Recently, sequencing-based approaches to microbial detection, based on targeted sequencing of the 16S ribosomal RNA gene and shotgun metagenomic sequencing, have been described for PJI diagnosis, with the former being more commonly used in current clinical practice [74,[80][81][82][83]. In 2022, bioMérieux received US Food and Drug Administration (FDA) authorization for the BioFire ® Joint Infection Panel, which interrogates a single synovial fluid sample against a 31 microbial target panel in approximately one hour. A limitation of panel-based diagnostics is that it only detects microorganisms that are included in the panel; for example, the aforementioned panel does not include S. epidermidis, an important PJI pathogen [82]. Novel microbial-based detection techniques remain an area of interest for rapid PJI diagnosis and microbial identification.
Similar to PJI, there are no perfect assays for NIAF diagnosis. Mechanical-related failures, such as aseptic loosening and fractures, are typically diagnosed and distinguished by X-ray, though non-mechanical failures, such as instability and adverse tissue reaction, may be difficult to differentiate from PJI due to inflammatory responses at affected areas [11,12,59,60]. It has been suggested that some non-mechanical-related NIAF cases may actually represent infection [196][197][198], although recent work using culture, PCR, and deep microbial sequencing has shown that microorganisms are rarely found in NIAF [73,81,[199][200][201]. Due to a lack of accurate NIAF diagnostic tools, a NIAF diagnosis may be contingent on the lack of a PJI diagnosis-that is, once arthroplasty failure is deemed to not be PJI-associated, NIAF is diagnosed. While helpful in determining whether antimicrobial treatment is necessary, the non-infectious pathogenesis underlying the failure may be unclear in such instances. Table 3. Sensitivity and specificity of synovial fluid C-reactive protein (CRP), calprotectin, interleukin-6 (Il-6), leukocyte esterase (LE), and lipocalin for periprosthetic joint infection diagnosis.

Importance of Fast and Accurate Arthroplasty Failure Diagnosis
The rise in arthroplasty procedures and the associated increase in PJI and NIAF, alongside the status of PJI diagnostics, justify the development of improved diagnostic techniques to differentiate PJI from NIAF and subsets within. Determining whether arthroplasty failure is due to PJI or NIAF and the causative organism, if infection is present, is important for selecting ideal management. As previously mentioned, results from currently used diagnostics may result in ambiguous classification, and in such cases, patients may receive unnecessarily broad-spectrum antimicrobial treatment or, alternatively, treatment that does not treat the actual cause of the arthroplasty failure. Unnecessary antimicrobial treatment may cause dysbiosis and drug-associated toxicity, in addition to aiding in the selection and expansion of antimicrobial-resistant bacteria, which is leading to dramatic effects on global health [53,202,203].
In 2022, the Antimicrobial Resistance Collaborators published an article estimating that 5 million deaths in 2019 worldwide were associated with bacterial antimicrobial resistance, although the full impact of antimicrobial resistance remains unknown due to the lack of global tracking systems. Methicillin-resistant S. aureus, an important cause of PJI, was a leading cause of death associated with antimicrobial resistance, resulting in more than 100,000 deaths in 2019 [204]. The World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC) have included antimicrobial-resistant S. aureus on their Priority Pathogen and Urgent Threat lists, respectively, due to its impact on global health [205,206]. S. epidermidis, an important PJI pathogen, is associated with high rates of methicillin resistance. The CDC reported a greater than 15% increase in antimicrobialresistant bacterial-associated infections and deaths in hospitals in 2020, possibly attributed to weakened infection prevention protocols and/or the usage of antibacterial agents during the COVID-19 pandemic [207].
These findings highlight the growing global crisis of antimicrobial resistance and illustrate the imperative of accurate and specific diagnoses of infectious diseases. Developing novel diagnostic techniques to differentiate PJI from NIAF and inform targeted antimicrobial usage will aid in patient management and in antimicrobial stewardship efforts, which will, in turn, assist in the fight against antimicrobial resistance globally.

Transcriptomic Immune Profiling
Transcriptomic analyses, such as RT-PCR and RNA-sequencing, have been conducted to interrogate the immune microenvironment during PJI and assess its potential impact on local bone and joint health. Transcriptomic studies performed on periprosthetic tissues from PJI show, unsurprisingly, that elevated expression of antimicrobial and immune cell activation genes dominates the immune response. In arthroplasty studies targeting specific transcripts, those associated with neutrophil activation, such as calprotectin, and IL-8, and macrophage inflammatory transcripts, such as chemokine (C-X-C motif) ligand (CXCL) 2 (CXCL2), and chemokine (C-C motif) ligand (CCL) 3 (CCL3), are elevated in PJI and associated with bone degeneration through bone-resorbing osteoclast generation, as well as induction of osteoblast inflammatory cytokine production [190,208,209]. Another targeted transcriptomic study of PJI-associated periprosthetic tissues found elevated levels of granulocyte colony-stimulating factor (G-CSF), IL-1β, IL-6, IL-8, and CD40L at infection sites [210]. In other targeted studies, levels of inflammatory mediators, such as toll-like receptor 2 (TLR-2), CCL2, presepsin, and osteopontin (OPN), were elevated in the serum of patients with PJI [186,211]. Of note, the last two are candidate biomarkers for sepsis diagnosis [212,213]. In addition to novel findings, many studies recapitulate antimicrobialrelated inflammatory biomarkers already used in PJI diagnosis, such as α-defensin, IL-6, and D-dimer.
Untargeted transcriptomic analysis of sonicate fluid from PJI patients has also been conducted. In a 2022 study by Masters et al., sonicate fluid from hip and knee arthroplasty failures underwent RNA sequencing; overall differentially expressed gene (DEG) profiling separated PJI and NIAF samples by principal component analysis (PCA). Pathway analysis found elevated DEGs to be primarily related to host defense, immune response, and cellular development and repair of canonical pathways. In all, 28 previously described potential PJI biomarkers and three novel potential biomarkers, including CCL20, coagulation factor F7, and Fc receptor-like 4 (FCRL4), were elevated in PJI compared to NIAF [214].

Proteomic Immune Profiling
Proteomic analyses of PJI and NIAF samples have also been conducted to assess the functional output of the immune response during arthroplasty failure and investigate its potential diagnostic use. Proteomic, similar to transcriptomic, profiling of arthroplasty failure, found that the local immune response during PJI is primarily driven by elevated antimicrobial inflammatory proteins, while the proteome of NIAF samples is more related to tissue homeostasis and wound healing. For example, targeted proteomic immunoassays conducted on synovial fluid have found neutrophil elastase (ELA-2), bactericidal/permeability-increasing protein (BPI), lipocalin, lactotransferrin, thrombospondin, IL-1β, IL-10, IL-1α, lactate, interferon (IFN)γ, IL-5, and IL-17A to be elevated in PJI compared to NIAF [102,107,147,184]. The diagnostic accuracy of α-defensin, CRP, IL-6, and LE were unsurprisingly recapitulated. Expression of the antimicrobial complement cascade of proteins has also been studied in synovial fluid using multiplex immunoassay. Complement proteins C1q, C3b/C3i, C4b, C5, C5a, MBL, and properdin were elevated in the PJI compared to NIAF. Individually, C1q was most able to differentiate PJI from NIAF, although the combination of elevated C1q, C3b/C3i, C4b, C5, C5a, and MBL was most predictive of PJI [215].

Cellular Immune Profiling
While the elevation of leukocytes, particularly neutrophils, is well-established in PJI, a robust understanding of the cellular profile during PJI is still being investigated. Cellularity profiling has primarily been conducted using synovial fluid and periprosthetic tissue. Due to the effects of processing, direct cellularity studies are limited when using sonicate fluid. To circumvent this, the transcriptomic results from the previously described Masters et al., 2022 bulk RNA-sequencing study on sonicate fluid were subjected to bioinformatic cellular deconvolution using CIBERSORTx [214,218]. Cellular deconvolution allows cellular analysis by "unmixing" bulk transcriptomic data to generate predicted cellularity profiles, in this case, made of 22 specific cell-types. Cellularity profiles created by CIBERSORTx are differentially clustered by PCA between PJI and NIAF. The differentiation of PJI and NIAF by predicted cellularity profiling was mainly separated by roles during inflammation-that is, cell types important for antimicrobial immunity were elevated in PJI, while NIAF populations were primarily composed of immune cells involved in tissue homeostasis and repair. In all, predicted total granulocyte, neutrophils, activated mast cells, CD8+ T cells, eosinophils, resting NK cells, activated CD4+ memory T cells were elevated in PJI, with predicted total macrophages/monocytes, M0 macrophages, M2 macrophages, total B cells, plasma cells, regulatory T cells, naïve B cells, and follicular helper T cells elevated in NIAF. Total granulocytes, neutrophils, and activated mast cells were most predictive of PJI from NIAF [218].
While it is known that infiltrating neutrophils are elevated during PJI, the role of mast cells during arthroplasty failure is uncharacterized. Tissue-resident mast cells have been described as "sentinel cells" able to detect microbial insults and initiate downstream antimicrobial inflammation by recruiting neutrophils and presenting bacterial antigens to the adaptive immune response, in additional to killing bacteria through secretion of antimicrobial peptides [219,220]. Activation of joint-specific mast cells has been linked to the induction of rheumatoid arthritis and increased joint inflammation during osteoarthritis and arthrofibrosis [174,[221][222][223]. As such, it could be reasoned that mast cells may play a role in antimicrobial host defense during PJI. Further investigation of the possible presence of this cell type in PJI is needed.
Findings from cellular deconvolution analysis of sonicate fluid have been largely recapitulated by results of flow cytometry experiments on synovial fluid and periprosthetic tissue. Anti-bacterial granulocytes, primarily driven by neutrophils and eosinophils, NK cells, and monocytes, were elevated in PJI vs. NIAF synovial fluid samples [224]. Similar to findings in synovial fluid, macrophages and monocytes were elevated in PJI-associated compared to NIAF periprosthetic tissues.
Although not normally characterized as an immune cell type, elevated platelets have been reported in the blood of patients with PJI; their diagnostic usefulness remains controversial [225][226][227]. There have been conflicting reports regarding the presence of T cells in arthroplasty failure, with findings ranging from lower or no to increased T cells in PJI compared to NIAF [210,224]. Whether, and if, T cells play a role during arthroplasty failure warrants future investigation. It has been suggested that the inflammatory response to bacterial biofilms during PJI may lead the recruitment of anti-inflammatory myeloidderived suppressor cells (MDSCs) to the joint, leading to immune response suppression and downstream chronic infection. Recruitment of MDSCs to the joint is elevated during PJI, likely due to the production of Il-12 by the host or of lactate and ATP synthase by certain biofilms themselves. Recruitment of MDSCs leads to the suppression of antimicrobial phagocyte recruitment and inhibits bacterial clearance [210,224,[228][229][230]. The role of MDSCs in arthroplasty failure is an area for future research.

Limitations of Immune Profiling for Arthroplasty Failure Diagnosis
In addition to the limitations of each individual technique to profile immune response during arthroplasty failure, there are limits associated with detailed immune profiling and its overall potential diagnostic use. At this time, virtually all studies profiling immune response during arthroplasty failure have been conducted as research studies. Though some results may portend future clinical use, when and whether these techniques can be validated for clinical use is unknown. Cohort sizes have been relatively small, with limited comorbidities addressed. As such, whether currently identified immune profiles will be recapitulated in larger, more diverse clinical populations is unknown. These studies have typically been conducted on samples from patients with clear PJI or NIAF diagnoses. It is unknown how these approaches will perform in more challenging to diagnose cases, the very cases where improved diagnostics are needed. The logistics of conducting multiomics analyses may also be problematic in clinical scenarios. Currently, these techniques and the necessary bioinformatic analyses are expensive and time-consuming. Prices of multi-omics analyses may decrease over time; the introduction of individualized medical tools, including personalized computational diagnostics, may become commonplace in the future.
A selection of cofactors-most not-yet-investigated-that may be relevant to accurate and reproducible results when assessing the immune response during arthroplasty failure is shown in Table 4. These variables may be important to note when planning, conducting, and analyzing the results of immune response profiling studies. Ways in which such variables impact immune responses to arthroplasty failure remain largely unexplored. As advanced multi-omic analyses develop, it will be interesting to investigate the clinical and diagnostic impact of these variables and their interactions.

The Future of PJI and NIAF Diagnostics
Multi-omics techniques to characterize immune response during arthroplasty failure represent a novel approach to potential future diagnosis of PJI and NIAF. While approaches described here differentiate PJI from NIAF, none have been discriminatory enough to define the underlying infectious organism(s) or cause of non-infectious failure, topics that deserve further study. Host-based diagnostics are not necessarily replacements for microbial-based detection tools, but will likely, instead, complement them. Complex bioinformatic tools, in combination with computational techniques, such as machine learning and artificial intelligence, are at the cutting edge of individualized diagnostics [231][232][233][234]. Advanced computational studies have already been conducted to better understand their potential use in PJI prediction [235][236][237][238]. These tools may be useful for addressing the current limitations of host-based profiling and synthesizing descriptive diagnostic readouts from multi-omics results in the future. Simultaneous assessment of host response and microorganism detection may allow single diagnostic reports to determine causative pathogen(s) and assess biological responses, providing insight into underlying key inflammatory etiologies and informing precision treatment ( Figure 5). Continued development of tools to detect causes of arthroplasty failure remains a challenge that warrants ongoing collaborative investigation.